Biostats: Rapid Review

Question 1

Type 1 error

Answer 1

False positive

Question 2

Type 2 error

Answer 2

False negative

Question 3

Incidence rate

Answer 3

of new cases over a certain time/total # of people at risk

Question 4

Prevalence rate

Answer 4

of active cases/total # of people at risk

Question 5

Sensitivity

Answer 5

True positive/(true positive + false negative)

Note: If you have the disease, what are the chances you will test positive.

Question 6

Specificity

Answer 6

True negative/(true negative + false positive)

Note: If you don’t have the disease, what are the chances you will test negative.

Question 7

Accuracy

Answer 7

(true positive + true negative)/total

Total: TP + TN + FP + FN

Question 8

Positive predictive value

Answer 8

True positive/(true positive + false positive)

Note: This is influenced by disease prevalence (if virtually no one has the disease its more likely to be a false positive than a true positive). If you test positive, what is your chance of actually having the disease.

Question 9

Negative predictive value

Answer 9

True negative/(true negative + false negative)

Note: This is influenced by disease prevalence. If you test negative, what is your chance of actually not having the disease.

Question 10

If you want to rule in a disease, you want a test with a high…

Answer 10

Specificity

Note: “SpIN” and “SnOUT”.

Question 11

How do you calculate absolute risk?

Answer 11

of new cases of disease/total # of people at risk

Note: This is the same way to calculate incidence rate.

Question 12

How do you calculate relative risk?

Answer 12

Incidence among people exposed/incidence among people not exposed

Question 13

How do you calculate odds ratio?

Answer 13

(cases with exposure x controls without exposure)/(controls with exposure x cases without exposure)

Note: This is usually used for retrospective studies.

Question 14

How does study power affect error rate?

Answer 14

Increasing study power reduces the chance of a type 2 error (false negative, not finding a difference when there actually is one)

Question 15

How do you calculate the number needed to treat?

Answer 15

1/absolute risk reduction

Question 16

How do you calculate absolute risk reduction?

Answer 16

(control event rate) - (experimental event rate)